Re: [PATCH v2 1/5] userfaultfd: introduce uffd_flags

From: Mike Rapoport <rppt@linux.ibm.com>
To: Nadav Amit <nadav.amit@gmail.com>
Cc: linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
	Nadav Amit <namit@vmware.com>,
	Mike Kravetz <mike.kravetz@oracle.com>,
	Hugh Dickins <hughd@google.com>,
	Axel Rasmussen <axelrasmussen@google.com>,
	Peter Xu <peterx@redhat.com>,
	David Hildenbrand <david@redhat.com>
Subject: Re: [PATCH v2 1/5] userfaultfd: introduce uffd_flags
Date: Sat, 23 Jul 2022 12:12:45 +0300	[thread overview]
Message-ID: <Ytu7jYc0jkWtHQV9@linux.ibm.com> (raw)
In-Reply-To: <20220718114748.2623-2-namit@vmware.com>

On Mon, Jul 18, 2022 at 04:47:44AM -0700, Nadav Amit wrote:
> From: Nadav Amit <namit@vmware.com>
> 
> As the next patches are going to introduce more information that needs
> to be propagated regarding handled user requests, introduce uffd_flags
> that would be used to propagate this information.
> 
> Remove the unused UFFD_FLAGS_SET to avoid confusion in the constant
> names.
> 
> Introducing uffd flags also allows to avoid mm/userfaultfd from being
> using uapi (e.g., UFFDIO_COPY_MODE_WP).
> 
> Cc: Mike Kravetz <mike.kravetz@oracle.com>
> Cc: Hugh Dickins <hughd@google.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Axel Rasmussen <axelrasmussen@google.com>
> Cc: Peter Xu <peterx@redhat.com>
> Cc: Mike Rapoport <rppt@linux.ibm.com>
> Acked-by: David Hildenbrand <david@redhat.com>
> Signed-off-by: Nadav Amit <namit@vmware.com>
> ---
>  fs/userfaultfd.c              | 22 +++++++++++---
>  include/linux/hugetlb.h       |  4 +--
>  include/linux/shmem_fs.h      |  8 +++--
>  include/linux/userfaultfd_k.h | 24 +++++++++------
>  mm/hugetlb.c                  |  3 +-
>  mm/shmem.c                    |  6 ++--
>  mm/userfaultfd.c              | 57 ++++++++++++++++++-----------------
>  7 files changed, 73 insertions(+), 51 deletions(-)
> 
> diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
> index e943370107d0..2ae24327beec 100644
> --- a/fs/userfaultfd.c
> +++ b/fs/userfaultfd.c
> @@ -1682,6 +1682,8 @@ static int userfaultfd_copy(struct userfaultfd_ctx *ctx,
>  	struct uffdio_copy uffdio_copy;
>  	struct uffdio_copy __user *user_uffdio_copy;
>  	struct userfaultfd_wake_range range;
> +	bool mode_wp;
> +	uffd_flags_t uffd_flags;
>  
>  	user_uffdio_copy = (struct uffdio_copy __user *) arg;
>  
> @@ -1708,10 +1710,15 @@ static int userfaultfd_copy(struct userfaultfd_ctx *ctx,
>  		goto out;
>  	if (uffdio_copy.mode & ~(UFFDIO_COPY_MODE_DONTWAKE|UFFDIO_COPY_MODE_WP))
>  		goto out;
> +
> +	mode_wp = uffdio_copy.mode & UFFDIO_COPY_MODE_WP;

This seems to be the only place where mode_wp is used in this function. 
I'd just drop it, and set uffd_flags directly from uffdio_copy.mode. E.g.
something like

	uffd_flags_t uffd_flags = UFFD_FLAGS_NONE;

	...

	if (uffdio_copy.mode & UFFDIO_COPY_MODE_WP)
		uffd_flags = UFFD_FLAGS_WP;

Otherwise

Acked-by: Mike Rapoport <rppt@linux.ibm.com>

> +
> +	uffd_flags = mode_wp ? UFFD_FLAGS_WP : UFFD_FLAGS_NONE;
> +
>  	if (mmget_not_zero(ctx->mm)) {
>  		ret = mcopy_atomic(ctx->mm, uffdio_copy.dst, uffdio_copy.src,
>  				   uffdio_copy.len, &ctx->mmap_changing,
> -				   uffdio_copy.mode);
> +				   uffd_flags);
>  		mmput(ctx->mm);
>  	} else {
>  		return -ESRCH;

-- 
Sincerely yours,
Mike.